Merge pull request #26 from philschmid/agent-example

Add Agent example & Fix stop sequences
philschmid · Aug 8, 2023 · 9147046 · 9147046
2 parents c2a77e5 + fbc1648
commit 9147046
Show file tree

Hide file tree

Showing 4 changed files with 364 additions and 11 deletions.
diff --git a/docs/examples/index.md b/docs/examples/index.md
@@ -4,14 +4,15 @@ Here are some examples to help you get started with the easyllm library:
 
 ## Hugging Face
 
-| Example                                                                 | Description                                                                           |
-| ----------------------------------------------------------------------- | ------------------------------------------------------------------------------------- |
-| [Detailed ChatCompletion Example](chat-completion-api)                  | Shows how to use the ChatCompletion API to have a conversational chat with the model. |
-| [Example how to stream chat requests](stream-chat-completions)          | Demonstrates streaming multiple chat requests to efficiently chat with the model.     |
-| [Example how to stream text requests](stream-text-completions)          | Shows how to stream multiple text completion requests.                                |
-| [Detailed Completion Example](text-completion-api)                      | Uses the TextCompletion API to generate text with the model.                          |
-| [Create Embeddings](get-embeddings)                                     | Embeds text into vector representations using the model.                              |
-| [Hugging Face Inference Endpoints Example](inference-endpoints-example) | Example on how to use custom endpoints, e.g. Inference Endpoints or localhost.        |
-| [Retrieval Augmented Generation using Llama 2](llama2-rag-example)      | Example on how to use Llama 2 70B for in-context retrival augmentation                |
+| Example                                                                 | Description                                                                            |
+| ----------------------------------------------------------------------- | -------------------------------------------------------------------------------------- |
+| [Detailed ChatCompletion Example](chat-completion-api)                  | Shows how to use the ChatCompletion API to have a conversational chat with the model.  |
+| [Example how to stream chat requests](stream-chat-completions)          | Demonstrates streaming multiple chat requests to efficiently chat with the model.      |
+| [Example how to stream text requests](stream-text-completions)          | Shows how to stream multiple text completion requests.                                 |
+| [Detailed Completion Example](text-completion-api)                      | Uses the TextCompletion API to generate text with the model.                           |
+| [Create Embeddings](get-embeddings)                                     | Embeds text into vector representations using the model.                               |
+| [Hugging Face Inference Endpoints Example](inference-endpoints-example) | Example on how to use custom endpoints, e.g. Inference Endpoints or localhost.         |
+| [Retrieval Augmented Generation using Llama 2](llama2-rag-example)      | Example on how to use Llama 2 70B for in-context retrival augmentation                 |
+| [Llama 2 70B Agent/Tool use example ](llama2-agent-example)             | Example on how to use Llama 2 70B to interace with tools and could be used as an agent |
 
 The examples cover the main functionality of the library - chat, text completion, and embeddings. Let me know if you would like me to modify or expand the index page in any way.
diff --git a/easyllm/clients/huggingface.py b/easyllm/clients/huggingface.py
@@ -159,7 +159,7 @@ def create(
         # create stop sequences
         if isinstance(request.stop, list):
             stop = stop_sequences + request.stop
-        if isinstance(request.stop, str):
+        elif isinstance(request.stop, str):
             stop = stop_sequences + [request.stop]
         else:
             stop = stop_sequences
@@ -348,7 +348,7 @@ def create(
         # create stop sequences
         if isinstance(request.stop, list):
             stop = stop_sequences + request.stop
-        if isinstance(request.stop, str):
+        elif isinstance(request.stop, str):
             stop = stop_sequences + [request.stop]
         else:
             stop = stop_sequences

diff --git a/mkdocs.yml b/mkdocs.yml
@@ -126,3 +126,4 @@ nav:
       - examples/get-embeddings.ipynb
       - examples/inference-endpoints-example.ipynb
       - examples/llama2-rag-example.ipynb
+      - examples/llama2-agent-example.ipynb